Skip to main content

Hi,

 

We are in the process of going live with MWO on Apps10 UPD 11 Oracle 12c and when adding 100 users have run into major performance bottlenecks with the MWO Batch Sync job. The job executes for days and there are numerous errors such as PlsqlapServer.ERRPING: No response from MWS server MAIN].

We have tried updating the number of batch processes to 3 and have not seen any performance improvements. Are they any customers currently using MWO on Apps10 upd 11 with Oracle 12 that have run into issues with the Batch sync job execution times and what did you do to see any improvements?

 

We have similar issues with the initialization which can take upwards of an hour.

 

Any inputs/suggestions are greatly appreciated.

 

If you increase no of process will make the situation worse becasue it will put the MWS in more unstable state. Need to address the ROOT cause. If a sync task takes more than 5 min to run, there there is a performance issue in the sync data process or data volume. Few things to check.

  • Have you added any Custom Fields to Sync entities? 
  • Have you applied all required Permission Set Filters to limit Data?
  • Any customizations in Sync entities? 

The error ‘No response from MWS server vMAIN]’ indirectly saying, it has already went on the peak. 

App10 Sync Troubleshooting guide has some information on this area: https://wit.ifsworld.com/f1docs/apps10upd7/Foundation1/040_administration/415_touch_apps/010_administration/070_trouble_shooting/default.htm


Thanks @kathlk. That’s exactly what happened, we reverted it back to 1. 2 entities take the longest one of them being the WorkObjectQuery which we’ve seen sync time of 1348526609770 which seems crazy. I looked at the query for this that is also using the mobile_sync_data_tab and this query at times is causing 100% CPU utilization on the database.

 

We are currently running the Bath Sync database task every 3 minutes and some of the entities (including WorkFlowObjectQuery) are set up to be daily on the sync rules. Is this set-up ideal?

 

Anything else we should be looking at?


I think you have a problem with the configuration.

Running Batch sync scheduled task in every 3 min can cause lot of issues. Do you know why you wanted to set that value? Default is 15 min and it runs in every 15 min and find pending entities based on sync shedule in the rules. Batch sync use to send basic data updates to device and it does not require a higher frequency. Push Data will send any operational data as it changed in the database. 

  • So, Set the batch sync back to 15 minute. 
  • Adjust Sync rules schedules to be run in off peak.
  • Check the sync volume and reduce record count using PS filters.

If you are in UPD12 (I think) or later, then Sync data volume screen available in the IEE. Otherwise you need to run a simple SQL on Mobile_Sync_Data view with group by App_Name, App_Version, User_Id, Device_Id will give you the record count on each entity per each user device. 


Thanks @kathlk. We have SSRS reports that get created after a survey is submitted and attached to the WO which we need to be synced to the app, which is why we had it at 3 minutes.

 

We spoke with IFS yesterday and they recommended the same thing as you to schedule the sync batch job at a freq of 15 minutes or higher. At this time, we are upgrading to UPD18 and checking if an upgrade to Oracle 19c along with changes to sync rules schedules and adding some permission sets to check if that helps.

 

 

 


Yes, Pls keep the sync shedule as 15 min, but if you need to sync an individual entity more frequently, then need to adjust the sync rule shedule of that entity. 

Hope this helps! 


Reply